Open file backup/sync --- like Outlook PST


#1

Hey guys… just wondering is it reasonable to be able to backup/sync files that are currently open? For instance if I have my Outlook pst file inside of Documents and I have Documents syncing to my “Documents Library” will that successfully preserve the PST file, or skip it?

Thanks!


#2

Whether or not you should open a file from a sync folder depends mostly upon how the software handles open files. Quickbooks, for example, has several files that update nearly constantly. So, whenever you open a data file from the sync drive, the sync program continually updates it on the cloud as well. It’s not a good idea to use the QB file directly from a syncing folder.

However, some software simply creates a lock file in the same folder, and it’s that file that’s under edit. Most Microsoft programs work that way and I would dare to say that Outlook may work the same way. I edit Excel files and Word files from synced folders frequently. It creates a lock file which gets synced. After you exit the program, the lock file goes away and therefore gets removed from the cloud. These files work quite well with syncing.

So, what I would do, is open up Outlook, check the folder and see if there is a lock file. Additionally, you want to watch what happens to it when you send or receive an email. If you have notifications turned on for your cloud service (I’m assuming Seafile), oh, it will let you know every time it updates. And, that can get annoying very quickly.


#3

Just did a little research. You can sync a PST file, but it’s not advisable, since most PST files are quite large. If you get a power interruption or network hiccup during a sync while editing, it could hose up the file.


#4

Agree! Syncing such big and often/randomly accessed files to another client, where these files are in use, too, can be very unreliable, even dangerous because of data loss.
But just using the sync feature of seafile to backup these files to the server should be ok.


#5

And that’s why I now use Thunderbird instead. Outlook lumps all the data into one large file. Thunderbird breaks every mailbox folder down into a separate file. It makes it a lot less likely for data corruption.


#6

Otherwise, if you want to backup a PST file, you should use a backup software instead of Seafile.
In our company we use Arx One Backup for our online backup… but it’s not a free tool.


#8

Thanks for your input guys… I kind of suspected that scenario, just wanted to confirm with others.

I don’t mind paying a reasonable fee for the software but I want a few features to be supported…

for Backend (destinatio):

  • SSH/Encryption of some sort
  • Incremental (rsync type ability to only update changes say in things like a PST file and not redo the entire thing each time)

Loving Seafile so far though as a self-hosted dropbox type replacement…


#9

You can encrypt libraries… But, you mention SSH, so I’m not certain what you are trying to achieve.

As for incremental updates to a file like a PST file, a PST file is a database. In order for Seafile or any other program to do that for that matter, it would have to have the ability to access those databases and manipulate them according to the specifications of the developer, in this case, Microsoft, while observing laws involved with manipulating intellectual property. Not only would it require a significant amount of code, it would also most likely require Seafile to utilize only those tools that Microsoft allows them to use. On top of that, Seafile would then have to chop the data up somehow in to smaller chunks for syncing. It isn’t feasible.

Thunderbird, however, operates entirely differently. It splits email folders into individual folders and each email is stored in its own file. So, it should lend itself well to being synced with Seafile.

I ditched Outlook several years ago because of the “Everything in one massive file” issue, long before I ever used Seafile. It’s a huge risk storing massive amounts of critical data in one file, and it’s why many database programs no longer do it, but rather split the data into blocks and such, in the same way Seafile does with its database.


#10

I completely agree on the Outlook issue… I don’t and haven’t used Outlooks for years (my environment is mainly Ubuntu based), but I have clients that I can’t get off Outlooks… that’s why I used that as an example.

Seafile will work great for syncing amongst multiple workstations, I’m also trying to find a stable solution for offsite backup (disaster recovery) … loving the idea of borg … but getting it to go on windows has been very challenging thus far.


#11

Yup… I have a couple of clients who insist on using Outlook. Don’t get me wrong… Outlook is a great program with some really nice features you won’t find elsewhere. It’s just their file structure that stinks. Ever tried to recover data from a large PST file? (They get large quickly). Yeah… fun… and gets a lot of people mad at you for not being able to recover everything.


#12

Lol… yep… that’s why I need a good backup solution! :slight_smile:


#13

@paul_d Take a look at Restic: https://restic.net , https://github.com/restic/restic
It is written in Go and compiled versions are available for Windows, too. It is a commandline tool. If you need a kind of GUI and scheduling out of the box, you can use Duplicati (https://www.duplicati.com). But I like the secure deduplication and snapshoting approve (a little bit like Seafiles filesystem) of Restic more!

I’m using Restic on my server to backup/snapshot Nextcloud, MariaDB, Gitlab, website document roots and a minecraft server to Backblaze B2 and so far I’m pretty impressed by this piece of software. In some areas it is not completely matured yet, but I think the core features are quite solid and the developer team especially the creator is supportive and very active and I find the community itself really nice.